EP0695445A1 - Method of estimating product distribution - Google Patents

Method of estimating product distribution

Info

Publication number
EP0695445A1
EP0695445A1 EP94912949A EP94912949A EP0695445A1 EP 0695445 A1 EP0695445 A1 EP 0695445A1 EP 94912949 A EP94912949 A EP 94912949A EP 94912949 A EP94912949 A EP 94912949A EP 0695445 A1 EP0695445 A1 EP 0695445A1
Authority
EP
European Patent Office
Prior art keywords
sales
prescription
locations
product
location
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Ceased
Application number
EP94912949A
Other languages
German (de)
French (fr)
Other versions
EP0695445A4 (en
Inventor
Mark Andrew Felthauser
Preston L. Mchenry
Harold Joseph Petrimoulx
Jeffrey Brian Schott
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
DUNS LICENSING ASSOCIATES LP
Original Assignee
DUNS LICENSING ASSOCIATES LP
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by DUNS LICENSING ASSOCIATES LP filed Critical DUNS LICENSING ASSOCIATES LP
Priority to EP06023574A priority Critical patent/EP1777654A3/en
Publication of EP0695445A1 publication Critical patent/EP0695445A1/en
Publication of EP0695445A4 publication Critical patent/EP0695445A4/en
Ceased legal-status Critical Current

Links

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06QINFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00Commerce
    • G06Q30/02Marketing; Price estimation or determination; Fundraising
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06QINFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00Administration; Management
    • G06Q10/10Office automation; Time management
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06QINFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00Commerce
    • G06Q30/02Marketing; Price estimation or determination; Fundraising
    • G06Q30/0201Market modelling; Market analysis; Collecting market data
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06QINFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00Commerce
    • G06Q30/02Marketing; Price estimation or determination; Fundraising
    • G06Q30/0201Market modelling; Market analysis; Collecting market data
    • G06Q30/0204Market segmentation
    • G06Q30/0205Location or geographical consideration
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06QINFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00Commerce
    • G06Q30/06Buying, selling or leasing transactions
    • GPHYSICS
    • G16INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16HHEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H20/00ICT specially adapted for therapies or health-improving plans, e.g. for handling prescriptions, for steering therapy or for monitoring patient compliance
    • G16H20/10ICT specially adapted for therapies or health-improving plans, e.g. for handling prescriptions, for steering therapy or for monitoring patient compliance relating to drugs or medications, e.g. for ensuring correct administration to patients

Definitions

  • the invention provides unbiased estimates of activity that are produced at the actual activity sources (i.e., each outlet and each prescriber at an outlet) and obviates the need for a predefined geographic area before estimates can be constructed.
  • nearest neighboring outlet activity sources within spatial proximity of each known but unsampled outlet are identified and the estimate of unsampled outlets is assembled by processing information on the discrete spatial correlation pattern among neighboring activity sources.
  • the spatial correlation process between outlets is represented by the distribution pattern of each prescriber's prescription activity among a local neighborhood of outlets (sampled or unsampled) .
  • activity estimates may be made for greater than 20,000 unsampled outlet locations and greater than 600,000 activity sources (e.g., prescribing physicians) using large quantities of sampled data (e.g., 2X10 9 sample records).
  • the distances between the selected sampled sales outlets and the unsampled sales outlet are combined with parameters characterizing each sales outlet to form a signal representing an estimate of sales of the product at the unsampled sales outlet.
  • product sales activity in an area including both sampled sales outlets and unsampled sales outlets are estimated by generating signals representative of the distances among the sales outlets, generating a signal representative of the type (e.g., size or sales volume) of each sales outlet, selecting a group of sampled sales outlets for each unsampled sales outlet, forming an estimate of the sales activity of each unsampled sales outlet according to the distances between the unsampled sales outlet and the associated selected sales outlet and the type representative signals for the unsampled and associated selected sampled sales outlets, and combining the sales activity data from the sampled sales outlets with the estimates of sales activity of the unsampled sales outlets.
  • the type e.g., size or sales volume
  • the prescription sales for a prescribing physician at a first prescription sales location in an area is estimated by correlating the prescription sales for the prescribing physician obtained from other prescription sales locations according to the distances between the first prescription sales location and selected ones of the other prescription sales locations.
  • the prescription sales activity for a prescribing physician at prescription sales outlets including sampled prescription sales outlets and unsampled prescription sales outlets are estimated by generating signals representing distances among the prescription sales outlets, generating signals representing the size or total volume of prescription sales at each prescription sales outlet, selecting a group of sampled outlets closest to each unsampled outlet and forming an estimate of prescription sales for the prescribing physician at each sampled outlet according to the distances between the sampled outlet and the unsampled outlets associated with the sampled outlet by the selection and the size or total volume of the outlets.
  • a signal representative of the prescription sales for the prescribing physician is produced according to the sales and estimates of sales for the prescribing physician at the sampled outlets.
  • d s is the distance between sampled pharmacy S and one of the other pharmacies U
  • T s is the total sales volume for all products at sampled pharmacy S
  • Tu is the total sales of all products at other pharmacy U
  • q is an number greater than 0
  • the summation ⁇ is over all sampled pharmacies S in the neighborhood N(U) .
  • the weighting factor signals, the sales data signals and the pharmacy characteristic signals are then transferred to a transfer store, divided therein and the divided portions are sent to a plurality of work station processors.
  • the volume of the particular product at each pharmacy U is then estimated in the work station processors as
  • a weighting factor for each sampled pharmacy in a neighborhood of an unsampled pharmacy is generated in the main processor and the sales data, the weighting factors and the pharmacy characteristics for the sampled pharmacies are transferred to a transfer store, divided therein and apportioned among the work station processors.
  • the sales volume of the prescription product V j for the prescribing physician j is estimated in the work station processors for the prescribing physician j according to
  • V sij is the prescription product sales volume for physician j at pharmacy i, w cauliflower, is the weighting factor generated in the main processor associating sampled pharmacy S to a neighborhood related unsampled pharmacy U
  • ⁇ ! is the summation over all sampled pharmacies
  • ⁇ 2 is the summation over weighting factors relating unsampled pharmacies p to sampled pharmacy Sij .
  • FIG. 3 is a flow chart showing the estimation of prescription sales at an outlet that is illustrative of the invention.
  • FIG. 4 is a flow chart showing the determination of sampled and unsampled outlets of FIG. 3 in greater detail
  • FIG. 7 is a flow chart showing the confidence signal operation of FIG. 6 in greater detail.
  • a single area 100 is shown in Fig. l for purposes of illustration only.
  • the estimation of sales activity at an unsampled outlet U is formed on the basis of the sales activity at the sampled outlets S in a neighborhood of the unsampled outlet U.
  • the neighborhood of an unsampled outlet U may be defined as the N closest sampled outlets S which is different for each unsampled outlet and not as a predefined geographic area.
  • the neighborhood of closest sampled outlets may all be located within a short distance of the unsampled outlet.
  • the neighborhood of N closest sampled outlets nay spread over distances. Consequently, each unsampled outlet has its own neighborhood area which varies according to the distances to the nearest N sampled outlets.
  • the correlation of sales activity data is not restricted to a predetermined geographic area as in the prior art.
  • FIG. 2 depicts a block diagram of the central station 120 of Fig. 1 which includes an input/output unit 201, a characteristics and location store 205, a sales data store 210, a processor 215, a program store 220, a bus 225, a transfer store 228 and work station processors 230-1 through 230-N.
  • Input/output 201 is coupled to bus 225 and is also coupled to the sampled outlet lines 130-1 to 130-N and to output line 245.
  • the characteristics and location store 205, the prescription data store 210 and the program store 225 are coupled to the bus 225.
  • the Processor 215 is coupled to the bus 220, the transfer store 228 and is also coupled to work station processors via a control line 240.
  • a data transfer from an outlet may occur for each sales transaction or may include a number of transactions for a prescribed period of time.
  • the sales data is sent to processor 215 and therein is evaluated in step 310 to determine the sampled outlets Si and the unsampled outlets Up in the processor 215.
  • Unsampled outlets may include outlets transferring data evaluated as invalid.
  • FIG. 4 shows a method of determining sampled and unsampled outlets in greater detail.
  • an outlet index n for the outlets 01, 02, ... , ON is set to 1 in step 401.
  • a sampled outlet index i and an unsampled outlet index p are set to 1 in steps 405 and 410.
  • the sales data for the particular product from each outlet On is checked in decision step 415 to determine if the data is valid (i.e., meets predetermined criteria) . If the data is judged to be valid in step 415, the outlet On is classified as a sampled outlet Si in step 420 and the index i is incremented in step 425.
  • step 415 When no data is available for the outlet On or the data is not accepted as valid in step 415, the outlet On is classified as an unsampled outlet Up in step 422 and the index p is incremented in step 427. The index n is then incremented in step 430. Until index n is greater than N for the last outlet ON, step 415 is reentered from decision step 435. When all of the outlets 01 through ON have been classified as sampled outlets and unsampled outlets, the last value of index p (pmax) and the last value of index i (imax) representing the number of unsampled outlets and the number of sampled outlets are stored in data store 210 (step 440) .
  • Fig. 1 there are five sampled outlets 110- 1 (01), 110-2 (02), 110-3 (03), 110-N-l (ON-1) and 110- N (ON) which are designated SI, S2, S3, S4 and S5 from the processing of FIG. 4 and two unsampled outlets 110- 4 (04) and 110-5 (05) which are designated as Ul and U2 from processing of FIG. 4.
  • Unsampled outlet Ul is located in the central portion of the area 100 and is surrounded by sampled outlets SI through S5.
  • Unsampled outlet U2 is located at one edge of the area 100, is closest to sampled outlet S5 and most remote from sampled outlet SI.
  • Priorly known techniques based an estimate of the sales volume of a product at an unsampled outlet on the sales volume of the product for the geographic area. Since the sales outlets have different characteristics (e.g., size and location) and have sales related to outlets outside a particular area, estimates based on the overall sales volume in a particular area as in the prior art are biased.
  • an estimate of sales volume of a particular product at a sales outlet is based on the known sales volume of other outlets according to the distances between the sales outlet and the other outlets and the particular characteristics of the outlets independent of any geographic area.
  • step 330 Signals corresponding to the distances between unsampled outlet Ul and sampled outlets SI through S5 and the distances between outlet U2 and sampled outlets SI through S5 are then formed in step 330.
  • step 335 the mmax closest sampled outlets to unsampled outlet p are selected. The selection is performed in the processor 215. mmax may be chosen according to the total number of sampled outlets. The selection of sampled outlets associated with each unsampled outlet is shown in greater detail in FIG. 5.
  • a set of distance signals dip' for sorting is generated in step 501 corresponding to the distance signals dip generated in step 330.
  • the unsampled outlet index p is set to one in step 505.
  • a selected outlet index m is set to one in step 510 and the sampled outlet index i is set to one in step 515.
  • a signal D is set to LPN (largest possible number) and the loop including steps 525, 530, 535 and 540 is entered to find the smallest distance of the distances dip' .
  • step 525 the signal dip' representing the distance from sampled outlet Si and unsampled outlet Up is compared to D.
  • D is set to dip'
  • Rm representing a tentative selected outlet is set to Si
  • the index i* is set to i
  • a tentative selected distance signal dmp is set to dip' in step 530.
  • Step 535 is then entered in which sampled outlet index i is incremented. Where dip' is not less than D, step 535 is entered directly from decision step 525.
  • Decision step 540 is then entered. Until sampled outlet index i exceeds imax in step 540, step 525 is reentered to compare the next distance signal dip' to the last determined minimum distance signal.
  • the minimum of the selected sampled outlets is chosen as Rm.
  • the minimum distance signal dip' is then set to LPN in step 545 to exclude Si* from comparison in step 525 and the selected outlet index m is incremented in step 550.
  • Step 515 is reentered from step 555 until mmax closest outlets for unsampled outlet p are selected and another outlet Rm is chosen in the loop from step 525 through 540.
  • the unsampled outlet index p is incremented in step 560 and step 510 is reentered via decision step 565 so that a set of m sampled outlets may be selected for the next unsampled outlet Up in FIG. 5 via decision step 565.
  • control is passed to step 340 in FIG. 3.
  • d ⁇ is the distance between sampled outlet Sm and unsampled outlet Up
  • T Up is the unsampled outlet characteristic (e.g., total sales volume for all products)
  • T m is the sampled outlet characteristic and q is greater than zero, q may, for example, be 2.
  • Index p is incremented in step 348 and control is passed to step 345 until p is greater than pmax in decision step 350.
  • the weighting factor signals for unsampled outlets Up and the product data for the outlets are read into the transfer store 228 as a data array which is divided therein into N data array portions.
  • the processor 215 sends control signals to work station processors 230-1 through 230-N to initiate processing of the data file portions in the work station processor.
  • Each work station processor then proceeds to form a product estimate signal for the data file portion assigned to it as indicated with respect to the entire data file in steps 355 through 370 in FIG. 3.
  • the loop from step 360-1 to step 370-1 is then entered.
  • the estimated sales of the particular product is then generated for a range of unsampled stores Up in step 360 according to
  • the unsampled outlet index p is incremented in step 365-1 and control is passed back to step 360-1 via decision step 370-1 until p is greater than the maximum of the range processed in work station processor 230-1 and an estimate of sales volume for all unsampled outlets in the range has been formed.
  • the processing of the other work station processors 230-2 through 230-N is the same as described with respect to the work station processor 230-1 except that the range is determined by the portion of the data file sent to the work station.
  • the processing in the work station processor 230-N is shown in the steps 355-N through 370-N.
  • Step 610 is entered from step 605 and signals representative of the distances dip between each sampled outlet Sij and each unsampled outlet Up are generated in the processor 215.
  • the set of nearest sampled pharmacy outlets Rmj for each unsampled pharmacy outlet Up is selected by processor 215 according to step 615. Selection of sampled pharmacies may be performed as described with respect to FIG. 5.
  • Each weighting factor signal is formed according to
  • T Up is the total sales volume of all products at pharmacy outlet Up
  • T ⁇ is the total sales volume of all products at pharmacy outlet Rm
  • a generalized variance function (GVF) is derived from the MSEs generated in step 701 of the form
  • log(SQRT(MSE) ) is then compared to K2 > Kl in step 735 to produce a medium estimated error signal in step 740 if log(SQRT(MSE) ) is less than K2. Where log(SQRT(MSE) ) is not less than K2, a high estimated error signal is generated in step 745. While three values for the estimated error signal are determined in the flow chart of FIG. 7, it is to be understood that any number of values such of 5 may be used.
  • the flow chart of FIG. 8 illustrates another arrangement for estimating the prescription product sales volume of a prescribing physician.
  • a group of sampled pharmacies is selected for each unsampled pharmacy and a weighting factor for each sampled pharmacy in the neighborhood of one of the unsampled pharmacies is generated as in Fig. 6.
  • the weighting factors for the sampled pharmacies are combined with the actual sales data for the sampled pharmacies according to
  • V ⁇ j ⁇ jV Sij [l + ⁇ 4Wip ]
  • Step 810 is entered from step 805 and signals representative of the distances dip between each sampled outlet Sij and each unsampled outlet Up are generated in the processor 215.
  • the set of nearest sampled pharmacy outlets R j for each unsampled pharmacy outlet Up is selected by processor 215 according to step 815.
  • the selection of sampled pharmacies Rmj may be performed as previously described with respect to FIG. 5.
  • V sij is the actual prescription product sales volume for physician j at pharmacy i
  • w ⁇ are the weighting factor for sampled pharmacy Sij
  • is the summation over all weighting factors associated with the pharmacy Sij.
  • [1 + ⁇ w, ⁇ ] is a projection factor for a physician's prescription at the sampled pharmacy V Slj .

Abstract

Sales activity of a product at sales outlets including sales outlets at which sales activity data is sampled (110-1 to 110-3) and unsampled sales outlets (110-4 to 110-5) is estimated by determining the distance between each of the sampled sales outlets and each of the unsampled sales outlets (step 330) and correlating sales activity data from the sampled sales outlets according to the determined distances. The sales activity volume of the product at the plurality of sampled outlets and the estimated sales activity volume of the product at the unsampled outlets are combined to obtain an estimate of sales activity for all the sales outlets (step 375). Sales activity of products prescribed by a physician at both the sampled and unsampled outlets can be estimated by correlating sales activity data for the prescribing physician at the sampled outlets according to the distances between the sampled outlets and the unsampled outlets.

Description

TITLE
METHOD OF ESTIMATING PRODUCT DISTRIBUTION
BACKGROUND OF THE INVENTION
Field of the Invention
The invention relates to marketing of products and, more particularly, to arrangements for estimating product distribution and sales. Background of the Invention
Manufacturers and distributors of retail products generally monitor product sales in order to maintain proper inventory and to direct marketing efforts. Monitoring may be done by sampling sales at retail outlets and transferring sales data to a central point for evaluation. Retail outlets usually cooperate in providing sales data but a significant number of retail outlets are not able to or do not elect to have sales data sampled in a form needed for analysis. As a result, it is necessary to estimate product sales of unsampled and poorly sampled individual outlets to provide marketing information.
In some industries, distribution of certain products is controlled so that additional data relating to the control of such products is needed for marketing purposes. In the pharmaceutical industry, for example, many products are sold by prescription and such sales are controlled by prescribing physicians. As a result, it is desirable to determine the number of prescriptions written by a physician so that marketing efforts may be directed at the prescribing physicians.
Estimates of business sales in small areas such as counties of a state have been made on the basis of known data for the state under the assumption that the relationships for the state also hold for the county. The article "Small-Area Estimation of Economic Statistics" by Cary T. Isaki, Journal of Business and Economic Statistics, Vol. D, No. 4, October, 1990, pages 435-441 describes a ratio correlation (multiple regression) approach for estimating retail sales for small areas (counties) using county-to-state shares of retail sales from two successive economic censuses. While these methods provide estimates of retail sales over a relatively small county area from publicly available data, they are not adapted to estimate retail sales of individual outlets where individual outlet characteristics differ widely. As a result, the estimates for individual outlets based on wide area data are biased and may not reflect actual sales of an individual outlet.
Estimation of physician prescribing activity has been attempted by marketing research practitioners based on ratio estimators and inflation factor estimators as commonly described in such texts as "Sampling Techniques" by W.G. Cochran, John Wiley, New York 1977. These methods attempt to estimate the activity in a pre-established geographic area of known dimensions by scaling up a sample of activity within the area in proportion to the level of a known auxiliary variable (i.e., ratio estimate) or in proportion to the level of sample coverage (via an inflation factor) for the entire area. Typical geographic areas encompass a plurality of outlets and prescribers. Such geographic- based methods do not yield estimates of each individual prescriber's activity within each individual outlet but only produce a measure of the total activity for the geography. If prescriber level estimates are desired, these methods must assume that the proportion of the total activity that is captured in the sample data (i.e., the captured proportion) of each prescriber is the same. If outlet estimates are desired, it must then be assumed that each unsampled outlet is accurately represented by the average of the sampled outlets in the geography. With these assumptions, all sample data within a stratum receive the same "scale- up" factor. These assumptions, however, are known to be false and result in biased estimates at the activity source level.
It is well known in the fields of resource exploration and mining to estimate data at locations which are not sampled and for which data is unavailable using data acquired at sampled locations by spatial correlation according to known physical principles. U.S. Patent 5,063,519 issued November 5, 1991 to Stanley W. Zison, for example, describes a method for estimating landfill gas production by measuring gas pressure in a soil cover at randomly selected locations in the landfill. A contour estimation of flow is generated and a prediction of the continuous spatial distribution of the landfill gas is produced.
Patent 4,995,011 issued February 19, 1991 to John L. Spiesberger describes an acoustic mapping system in which the positions of acoustic transmitters is determined by processing data from five or more receivers distributed in a marine or terrestrial environment. The processing includes cross correlation of the receiver data and provides displays of computer generated maps.
U.S. Patent 4,926,394 issued May 15, 1990 to Philip M. Doyen discloses a Monte Carlo method for estimating lithography from seismic data in which discrete geological measurements of rock properties are combined with continuous measurements of seismic attributes. The combined measurements are converted into a display of the best estimate of subsurface rock classes.
The aforementioned patents are directed to spatially continuous processes in which estimation is done either by interpolating continuous spatial processes from data at known locations or by Monte Carlo simulations using the known data to determine an optimized estimate. In each process, reliance is placed on the laws of the underlying continuous physical process. It would be desirable to use spatial correlation to provide estimates of retail sales. Sales outlets such as pharmacies, however, do not form a group to which a common physical process applies but each outlet is a separate discrete entity with its own characteristics. Consequently, there is no underlying continuous physical process on which to base such correlation as in the aforementioned patents.
Brief Summary of the Invention
The invention is directed to estimating sales activity of a product by determining the distances between first sales outlets from which sales activity data are obtained and other sales outlets and correlating the sales activity data of the first sales outlets according to the determined distances between the first and the other sales outlets.
In contrast to priorly known arrangements, the invention provides unbiased estimates of activity that are produced at the actual activity sources (i.e., each outlet and each prescriber at an outlet) and obviates the need for a predefined geographic area before estimates can be constructed. According to the invention, nearest neighboring outlet activity sources within spatial proximity of each known but unsampled outlet are identified and the estimate of unsampled outlets is assembled by processing information on the discrete spatial correlation pattern among neighboring activity sources. The spatial correlation process between outlets is represented by the distribution pattern of each prescriber's prescription activity among a local neighborhood of outlets (sampled or unsampled) . In practice, activity estimates may be made for greater than 20,000 unsampled outlet locations and greater than 600,000 activity sources (e.g., prescribing physicians) using large quantities of sampled data (e.g., 2X109 sample records).
According to one aspect of the invention, a distance between one sales outlet at which sales activity is not sampled and each of a plurality of sales outlets at which sales activity is sampled is determined. A group of the sampled sales outlets in a neighborhood of the one unsampled sales outlet is selected and sales activity data of the product at the selected sales outlet are combined according to the distances between the one unsampled sales outlet and the selected sales outlets to form an estimate of the sales at the unsampled sales outlet.
According to another aspect of the invention, the distances between the selected sampled sales outlets and the unsampled sales outlet are combined with parameters characterizing each sales outlet to form a signal representing an estimate of sales of the product at the unsampled sales outlet.
According to yet another aspect of the invention, product sales activity in an area including both sampled sales outlets and unsampled sales outlets are estimated by generating signals representative of the distances among the sales outlets, generating a signal representative of the type (e.g., size or sales volume) of each sales outlet, selecting a group of sampled sales outlets for each unsampled sales outlet, forming an estimate of the sales activity of each unsampled sales outlet according to the distances between the unsampled sales outlet and the associated selected sales outlet and the type representative signals for the unsampled and associated selected sampled sales outlets, and combining the sales activity data from the sampled sales outlets with the estimates of sales activity of the unsampled sales outlets.
According to yet another aspect of the invention, the prescription sales for a prescribing physician at a first prescription sales location in an area is estimated by correlating the prescription sales for the prescribing physician obtained from other prescription sales locations according to the distances between the first prescription sales location and selected ones of the other prescription sales locations.
According to yet another aspect of the invention, the prescription sales activity for a prescribing physician in an area including sampled prescription sales outlets and unsampled prescription sales outlets are estimated by generating signals representing distances among the prescription sales outlets, generating signals representing the size or total volume of prescription sales at each prescription sales outlet, selecting a group of sampled outlets closest to each unsampled outlet and forming an estimate of prescription sales for the prescribing physician at each unsampled outlet according to the distances between the unsampled outlet and the selected sampled outlets and the size or total volume of the outlets. A signal representative of the prescription sales for the prescribing physician is produced according to the sales for the prescribing physician at the sampled outlets and the estimates of the sales for the prescribing physician at the unsampled outlets.
According to yet another aspect of the invention, the prescription sales activity for a prescribing physician at prescription sales outlets including sampled prescription sales outlets and unsampled prescription sales outlets are estimated by generating signals representing distances among the prescription sales outlets, generating signals representing the size or total volume of prescription sales at each prescription sales outlet, selecting a group of sampled outlets closest to each unsampled outlet and forming an estimate of prescription sales for the prescribing physician at each sampled outlet according to the distances between the sampled outlet and the unsampled outlets associated with the sampled outlet by the selection and the size or total volume of the outlets. A signal representative of the prescription sales for the prescribing physician is produced according to the sales and estimates of sales for the prescribing physician at the sampled outlets.
In one embodiment of the invention, sales of a particular product are sampled at a first group of the pharmacies and are sent to a central station having a main processor and a group of work station processors. In the main processor, data representing the distances between the first group of pharmacies and each other pharmacy are generated and a signal representative of the total sales of all products of each pharmacy or the size of each pharmacy. A weighting factor for the sales of the particular product at each of the sampled pharmacies S in a neighborhood N(U) of the other pharmacy U is generated according to
ws = { ( l/ds) / (∑τs/d§) } *τυ
where ds is the distance between sampled pharmacy S and one of the other pharmacies U, Ts is the total sales volume for all products at sampled pharmacy S, Tu is the total sales of all products at other pharmacy U, q is an number greater than 0 and the summation Σ is over all sampled pharmacies S in the neighborhood N(U) . The weighting factor signals, the sales data signals and the pharmacy characteristic signals are then transferred to a transfer store, divided therein and the divided portions are sent to a plurality of work station processors. The volume of the particular product at each pharmacy U is then estimated in the work station processors as
V„ = EWsVvs,
where Vs is the sales volume of the particular product at a selected sampled pharmacy, ws is the weighting factor for the selected sampled pharmacy and the summation Σ is over all sampled pharmacies S in the neighborhood N(U) . In another embodiment of the invention, sales of a prescription product of a prescribing physician j are sampled at a first group of the pharmacies and are sent to a central station having a main processor and plural work station processors. At the main processor of the central station, data representing the distances dsu between the each of selected pharmacies S of the first group of pharmacies and each other pharmacy U is generated and a signal representative of the total sales of each pharmacy Ts and Tu is stored. A weighting factor for each sampled pharmacy in a neighborhood of an unsampled pharmacy is generated in the main processor and the sales data, the weighting factors and the pharmacy characteristics for the sampled pharmacies are transferred to a transfer store, divided therein and apportioned among the work station processors. The sales volume of the prescription product Vj for the prescribing physician j is estimated in the work station processors for the prescribing physician j according to
VTj = Σ.Vsii [l + Σ2w
where Vsij is the prescription product sales volume for physician j at pharmacy i, w„, is the weighting factor generated in the main processor associating sampled pharmacy S to a neighborhood related unsampled pharmacy U, ∑! is the summation over all sampled pharmacies and Σ2 is the summation over weighting factors relating unsampled pharmacies p to sampled pharmacy Sij .
wsu = {(1/d.S )/(Σ3Ts/d£)}*Tu
where Σ3 is the summation of over sampled pharmacies in a neighborhood of an unsampled pharmacy. The features and advantages of the present invention will become apparent from the following detailed description of preferred embodiments thereof taken together with the accompanying drawings.
Description of the Drawings
FIG. 1 is a block diagram showing an arrangement of sales outlets and a processing station illustrative of an embodiment of the invention;
FIG. 2 is a block diagram of a processing system that may be used as the central station of FIG. 1;
FIG. 3 is a flow chart showing the estimation of prescription sales at an outlet that is illustrative of the invention;
FIG. 4 is a flow chart showing the determination of sampled and unsampled outlets of FIG. 3 in greater detail;
FIG. 5 is a flow chart showing the selection of the group of sampled outlets for each unsampled outlet in greater detail;
FIG. 6 is a flow chart showing one arrangement for the estimation of prescriptions sales for a prescribing physician at a plurality of pharmacies illustrative of the invention;
FIG. 7 is a flow chart showing the confidence signal operation of FIG. 6 in greater detail; and
FIG. 8 is a flow chart showing another arrangement for the estimation of prescription sales for a prescribing physician at a plurality of pharmacies illustrative of the invention. Detailed Description
FIG. 1 depicts an arrangement illustrating a first embodiment of the invention in which product sales at unsampled sales outlets are estimated. In FIG. 1, there are shown an area 100, sampled sales outlets 110- 1, 110-2, 110-3, 110-N-l and 110-N and unsampled sales outlets 110-4 and 110-5 in the area 100 and a central station 120. Each of sampled outlets 110-1, 110-2, 110-3, 110-N-l and 110-N may preferably be coupled via a line of lines 130-1 through 130-N to the central station 120.
In FIG. 1, the outlets may be pharmacies or other type of retail stores or distribution establishments all of which distribute a particular product. The outlets are at various locations in the area 100. While there are 7 outlets shown in FIG. 1 for purposes of illustration, it is to be understood that there are generally hundreds or thousands of outlets which are not restricted to a given area. The location of each outlet is generally known in terms of latitude and longitude from available census data or in terms of zip code centroids from Post Office data. Accordingly, the distances between pharmacies can be determined.
Product sales data generated at each outlet Sn (e.g., 110-1) is preferably transferred to the central station 120 via a line (e.g., 130-1). Unsampled outlets Un (e.g., 110-4 or 110-5) in the area 100 are not coupled to the central station or, if coupled, do not supply valid sales data so that only an estimate of the sales volume of the particular product can be made.
A single area 100 is shown in Fig. l for purposes of illustration only. According to the invention, the estimation of sales activity at an unsampled outlet U is formed on the basis of the sales activity at the sampled outlets S in a neighborhood of the unsampled outlet U. The neighborhood of an unsampled outlet U may be defined as the N closest sampled outlets S which is different for each unsampled outlet and not as a predefined geographic area. In an urban area, the neighborhood of closest sampled outlets may all be located within a short distance of the unsampled outlet. In a rural area, the neighborhood of N closest sampled outlets nay spread over distances. Consequently, each unsampled outlet has its own neighborhood area which varies according to the distances to the nearest N sampled outlets. Advantageously, the correlation of sales activity data is not restricted to a predetermined geographic area as in the prior art.
FIG. 2 depicts a block diagram of the central station 120 of Fig. 1 which includes an input/output unit 201, a characteristics and location store 205, a sales data store 210, a processor 215, a program store 220, a bus 225, a transfer store 228 and work station processors 230-1 through 230-N. Input/output 201 is coupled to bus 225 and is also coupled to the sampled outlet lines 130-1 to 130-N and to output line 245. The characteristics and location store 205, the prescription data store 210 and the program store 225 are coupled to the bus 225. The Processor 215 is coupled to the bus 220, the transfer store 228 and is also coupled to work station processors via a control line 240. The transfer store 228 is coupled to work station processors 230-1 to 230-N and the work station processors are coupled together through a network 250. Characteristics signals stored in store 205 may include signals representing the type of outlet or the total sales of all products at the outlet. This type of data is available from sources such as Drug Distribution Data (DDD ™) available from IMS America, Plymouth Meeting, Pennsylvania. DDD ™ is a trademark of IMS America. Program store 220 stores instruction signals that control the operation of the processor 215 and provide parameter signals to determine the operation of work station processors 230-1 through 230-N through the processor 215 and control line 240.
The sales outlet data received by input/output 201 from sales outlets which may exceed 2X109 records each having between 88 and 1000 bytes is transferred to data store 210. In view of the large amount of data to be processed, the processing is divided between the processor 215 and the work station processors 230-1 through 230-N. Information signal arrays produced in the processor 215 are transferred to work station processors 230-1 through 230-N through the transfer store 228. Each information signal array from the processor 215 placed in the transfer store 228 is divided into N portions. A preassigned portion of the information signal array in the transfer store is supplied to each of work station processors 230-1 through 230-N and the processing of the portions in the work stations 230-1 through 230-N is controlled by signals from the processor 215 via the control line 240. After the processing of the information signal array portions in the work station processors, the processed information signal array portions are merged into one processed information signal array which is returned to transfer store 228 from the work station processors. The returned information signal array in the transfer store 228 is then further processed in the processor 215 to produce estimate sales volume results. The operation of the system of FIG. 2 will be further described in connection with the estimation arrangements shown in FIGs. 3-7.
FIG. 3 depicts a flow chart illustrating the operation of the central station 120 of Fig. 1 in estimating the volume of sales of a particular product at an unsampled outlet such as outlet 04 or outlet 05 in Fig. 1. The operations depicted in the flow chart of FIG. 3 are performed by processor 215 and work station processors 230-1 to 230-N of FIG. 2 under control of instruction signals from in the program store 220. In the flow chart of FIG. 3, product data from outlets 01, 02, 03, ON-1 and ON are transferred to the input/output unit 201 preferably via corresponding lines 130-1 through 130-N in step 301. Transferred data is stored in outlet sales data store 210. A data transfer from an outlet may occur for each sales transaction or may include a number of transactions for a prescribed period of time. At preset intervals, the sales data is sent to processor 215 and therein is evaluated in step 310 to determine the sampled outlets Si and the unsampled outlets Up in the processor 215. Unsampled outlets may include outlets transferring data evaluated as invalid.
FIG. 4 shows a method of determining sampled and unsampled outlets in greater detail. Referring to FIG. 4, an outlet index n for the outlets 01, 02, ... , ON is set to 1 in step 401. A sampled outlet index i and an unsampled outlet index p are set to 1 in steps 405 and 410. The sales data for the particular product from each outlet On is checked in decision step 415 to determine if the data is valid (i.e., meets predetermined criteria) . If the data is judged to be valid in step 415, the outlet On is classified as a sampled outlet Si in step 420 and the index i is incremented in step 425. When no data is available for the outlet On or the data is not accepted as valid in step 415, the outlet On is classified as an unsampled outlet Up in step 422 and the index p is incremented in step 427. The index n is then incremented in step 430. Until index n is greater than N for the last outlet ON, step 415 is reentered from decision step 435. When all of the outlets 01 through ON have been classified as sampled outlets and unsampled outlets, the last value of index p (pmax) and the last value of index i (imax) representing the number of unsampled outlets and the number of sampled outlets are stored in data store 210 (step 440) .
As shown in Fig. 1, there are five sampled outlets 110- 1 (01), 110-2 (02), 110-3 (03), 110-N-l (ON-1) and 110- N (ON) which are designated SI, S2, S3, S4 and S5 from the processing of FIG. 4 and two unsampled outlets 110- 4 (04) and 110-5 (05) which are designated as Ul and U2 from processing of FIG. 4. Unsampled outlet Ul is located in the central portion of the area 100 and is surrounded by sampled outlets SI through S5. Unsampled outlet U2 is located at one edge of the area 100, is closest to sampled outlet S5 and most remote from sampled outlet SI. Priorly known techniques based an estimate of the sales volume of a product at an unsampled outlet on the sales volume of the product for the geographic area. Since the sales outlets have different characteristics (e.g., size and location) and have sales related to outlets outside a particular area, estimates based on the overall sales volume in a particular area as in the prior art are biased. In accordance with the invention, an estimate of sales volume of a particular product at a sales outlet is based on the known sales volume of other outlets according to the distances between the sales outlet and the other outlets and the particular characteristics of the outlets independent of any geographic area. By using the outlet characteristics and the distances, an unbiased and more accurate estimate may be determined.
Signals corresponding to the distances between unsampled outlet Ul and sampled outlets SI through S5 and the distances between outlet U2 and sampled outlets SI through S5 are then formed in step 330. In step 335, the mmax closest sampled outlets to unsampled outlet p are selected. The selection is performed in the processor 215. mmax may be chosen according to the total number of sampled outlets. The selection of sampled outlets associated with each unsampled outlet is shown in greater detail in FIG. 5.
With reference to FIG. 5, a set of distance signals dip' for sorting is generated in step 501 corresponding to the distance signals dip generated in step 330. The unsampled outlet index p is set to one in step 505. A selected outlet index m is set to one in step 510 and the sampled outlet index i is set to one in step 515. In step 520, a signal D is set to LPN (largest possible number) and the loop including steps 525, 530, 535 and 540 is entered to find the smallest distance of the distances dip' .
In decision step 525, the signal dip' representing the distance from sampled outlet Si and unsampled outlet Up is compared to D. When dip' is less than D, D is set to dip' , Rm representing a tentative selected outlet is set to Si, the index i* is set to i and a tentative selected distance signal dmp is set to dip' in step 530. Step 535 is then entered in which sampled outlet index i is incremented. Where dip' is not less than D, step 535 is entered directly from decision step 525. Decision step 540 is then entered. Until sampled outlet index i exceeds imax in step 540, step 525 is reentered to compare the next distance signal dip' to the last determined minimum distance signal. When i exceeds imax, the minimum of the selected sampled outlets is chosen as Rm. The minimum distance signal dip' is then set to LPN in step 545 to exclude Si* from comparison in step 525 and the selected outlet index m is incremented in step 550.
Step 515 is reentered from step 555 until mmax closest outlets for unsampled outlet p are selected and another outlet Rm is chosen in the loop from step 525 through 540. Upon selection of mmax sampled outlets, the unsampled outlet index p is incremented in step 560 and step 510 is reentered via decision step 565 so that a set of m sampled outlets may be selected for the next unsampled outlet Up in FIG. 5 via decision step 565. When p is greater than pmax, control is passed to step 340 in FIG. 3.
In step 340, index p is set to one. A weighting factor wm is then determined for each selected sampled outlet Rm of unsampled outlet Up in step 345. Weighting factor generation is performed in the processor 215. The weighting factor is an inverse function of the distance between the sampled outlet Rm and the unsampled outlet Up and the characteristics of the sampled and unsampled outlets according to
W m = { ( 1/dUu / (∑ ITm/ Om-v} } *T Up
where d^^ is the distance between sampled outlet Sm and unsampled outlet Up, the summation is over all sampled outlets for m=l to mmax, TUp is the unsampled outlet characteristic (e.g., total sales volume for all products) , Tm is the sampled outlet characteristic and q is greater than zero, q may, for example, be 2. Index p is incremented in step 348 and control is passed to step 345 until p is greater than pmax in decision step 350.
The weighting factor signals for unsampled outlets Up and the product data for the outlets are read into the transfer store 228 as a data array which is divided therein into N data array portions. The processor 215 sends control signals to work station processors 230-1 through 230-N to initiate processing of the data file portions in the work station processor. Each work station processor then proceeds to form a product estimate signal for the data file portion assigned to it as indicated with respect to the entire data file in steps 355 through 370 in FIG. 3. In step 355-1, a starting value of the unsampled outlet index p=l is set. The loop from step 360-1 to step 370-1 is then entered. The estimated sales of the particular product is then generated for a range of unsampled stores Up in step 360 according to
Est(vUp) = rwmvβ
where Vm is the sales volume of the particular product at sampled outlet m and the summation is over the sampled outlets from m=l to m=mmax. The unsampled outlet index p is incremented in step 365-1 and control is passed back to step 360-1 via decision step 370-1 until p is greater than the maximum of the range processed in work station processor 230-1 and an estimate of sales volume for all unsampled outlets in the range has been formed. The processing of the other work station processors 230-2 through 230-N is the same as described with respect to the work station processor 230-1 except that the range is determined by the portion of the data file sent to the work station. The processing in the work station processor 230-N is shown in the steps 355-N through 370-N.
For purposes of illustration with respect to Fig. l, the number of selected sampled outlets mmax is chosen as 3. It is to be understood, however, that other values may be chosen. For example, if there are 50 or more sampled outlets, mmax = 10 is a suitable value. In Fig. l, sampled outlets 01, 02 and 03 are selected as the three closest sampled outlets Rl, R2 and R3 to unsampled outlet Ul. To illustrate the invention, assume that the distance dRIU1 from sampled outlet Rl to unsampled outlet Ul is 0.4 miles, the distance d^u, between sampled outlet R2 and unsampled outlet Ul is . 0.2 miles and the distance d^,., between sampled outlet R3 and unsampled outlet Ul is 0.6 miles. Further assume that the total sales volume for all products at sampled outlets Rl, R2, R3 and Ul are $3,000, $2,000, $5,000 and $4,000, respectively. The weighting factor for sampled outlet R2 is then
w2 = { (1/0.2)7(2000/(0.2)2 + 3000/(0.4)2 + 5000/(0.6)2) }*4000
w2 = 1.210084
Similarly, j = 0.302521 and w3 = 0.13445377. For a sales volume of the particular product at Rl, R2 and R3 of 5, 20 and 4, respectively, the estimated sales volume of the particular product at unsampled outlet Ul is
Est (V ) = !*VRj + W2VR2 + WjVna
Est (Vul ) = 26 .252
The product volume signals for the sampled outlets Si is then formed in step 372-1 through 372-N and the total estimated sales volume of the product for unsampled and the sampled outlets is then formed for each range in the work station processors 230-1 through 230-N in steps 375. The resulting unsampled outlet estimate and total volume estimate signals of the processing in the work station processors is then merged in and totalled step 380 into a result data file. The result data file is transferred to transfer store 228 and therefrom to data store 210. The results are then sent to output line 245 of the input/output 201.
FIG. 6 shows a flow chart illustrating estimation of the distribution of a controlled product by a control authority. More particularly, FIG. 6 shows the operation of the arrangement of FIGS, l and 2 in estimating sales of a prescription product for a prescribing physician at the pharmacies. The operations in the flow chart of FIG. 6 are performed by processor 215 and work station processors 230-1 through 230-N of FIG. 2 under control of corresponding instruction signals stored in the program store 220. Referring to FIG. 6, location data, data of type characteristics of pharmacy outlets 01 through ON in FIG. 1 and physician identification data are stored in the characteristics and locations store 205 of FIG. 2 in step 601. Prescription data is transferred from pharmacies 01, 02, 03, ON-1 and ON and is stored in pharmacy outlet data store 210 according to the prescribing physicians (step 603) .
At prescribed intervals, the total sales for the prescribing physician j of a particular prescription product is estimated in steps 605 through 670 of FIG. 6. In step 605, the processor 215 operates to determine the sampled pharmacy outlets Sij and the unsampled pharmacy outlets Up for a particular prescription product according to the validity and volume of the transferred prescription data of the prescribing physician j. The arrangement shown in FIG. 4 may be used in the determination of step 605. As described with respect to FIGS. l and 3, pharmacy outlets 01, 02, 03, ON-1 and ON can be determined as sampled outlets Slj , S2j , S3j , S4j and S5j where the sampled data is validated. Outlets 04 and 05 are classified as unsampled outlets Ul and U2.
Step 610 is entered from step 605 and signals representative of the distances dip between each sampled outlet Sij and each unsampled outlet Up are generated in the processor 215. After the distance determination of step 610, the set of nearest sampled pharmacy outlets Rmj for each unsampled pharmacy outlet Up is selected by processor 215 according to step 615. Selection of sampled pharmacies may be performed as described with respect to FIG. 5. Then, the unsampled pharmacy outlet index p is set to 1 in step 620 and a weighting factor signal wm for the each sampled pharmacy outlet Rm (m=l to mmax) is generated in loop from step 625 to step 634. Each weighting factor signal is formed according to
wm = { (1/d j)/(∑( d^ }*TUp
where q is greater than O, TUp is the total sales volume of all products at pharmacy outlet Up, T^ is the total sales volume of all products at pharmacy outlet Rm and the summation Σ is from m=l to mmax.
After the weighting factor signals have been formed for the last unsampled pharmacy in the loop from step 625 to 634, a data array including the weighting information and the sales data from store 210 is formed by the processor 215 and sent to transfer store 228. The data array is divided into N portions each of which is processed by one of work station processors 230-1 through 230-N to form a signal representing an estimate of the total prescription product sales volume for physician j . The operations of the work station processors are controlled by the processor 215 through the control line 245. Each work station processor operates to process a predetermined range of the data array in the transfer store 228.
The work station processor 230-1 operates according to steps 638-1 through 655-1 to form the volume product signals for physician j in the range p=l to p=pl. In step 640-1, the prescription estimate signal for unsampled outlets Upj is formed for each unsampled pharmacy in the range from p=l to p=pl and the sales volume signal for the sampled pharmacies in this range is determined by the work station processor 230-1 according to step 655-1. Work station processor 230-N operates in similar manner for physician j over the range p=pN to p=pmax as indicated in FIG. 6 according to steps 638-N through 655-N. Signals are transferred from one work station processor to another as required for the operation of the one work station processor through the network 255. The results of the operation of work station processors 230-1 through 230-N are merged in step 660 and an estimate of the prescription product volume
Vτ = Σ,Vsij + Σ2VjUp
is generated in step 665 where
VjUp = Σ3 { { ( l/d up ) / (3 ( j,) } *τUp}vRmj
Σ, is the summation over all sampled outlets, Σ2 is the summation over all unsampled outlets, Σ3 is the summation over all sampled outlets in the neighborhood of unsampled outlet Up. At this time, a confidence signal that estimates the degree of possible error of the total volume VTj of the product prescriptions of the prescribing physician j is then generated in step 670. FIG. 7 shows the confidence signal generation operation of step 670 in greater detail. Referring to FIG. 7, a mean squared error signal MSE is first generated by bootstrapping on the basis of the sales data from the sampled pharmacy outlets SI through Simax in step 701. The bootstrapping method is well known in the art and is described in "The Jacknife, The Bootstrap, and Other Resampling Plans" by B. Efron, Society for Industrial and Applied Mathematics (SIAM) Publications, Philadelphia 1982.
In the bootstrapping, subsets of pharmacy outlets are selected and the prescribing physician's prescription volume is estimated therefrom. The variances of the "bootstrapped" estimates closely approximates the true variance. A generalized variance function (GVF) is derived from the MSEs generated in step 701 of the form
log(SQRT(MSE) ) = a + blog(Tj) + clog(Nj) + dlog(Nsj)
where SQRT is the square root, a, b, c and d are regression coefficients, Tj is the estimated total of the prescription product prescribed by physician j , N. is the total of prescription products prescribed by physician j and dispensed at the sampled pharmacy outlets and Nsj is the number of sampled pharmacy outlets with prescription product sales for physician j . The generalized variance function is described in "Introduction to Variance Estimation" by K.M. Wolter, Springer-Verlag, New York 1985.
The values of Tjf Nj and NSj are determined in steps 710, 715 and 720 from the prescription data in store 210 of FIG. 2 in processor 215. Regression coefficient signals a, b, c and d are generated by multiple regression techniques well known in the art and a log(SQRT(MSE) ) value for the physician j is determined in step 720. Decision step 725 is then entered in which the value log(SQRT(MSE) ) is compared to Kl. If log(SQRT(MSE) ) is less than Kl, a low estimated error signal is produced in step 730. The value log(SQRT(MSE) ) is then compared to K2 > Kl in step 735 to produce a medium estimated error signal in step 740 if log(SQRT(MSE) ) is less than K2. Where log(SQRT(MSE) ) is not less than K2, a high estimated error signal is generated in step 745. While three values for the estimated error signal are determined in the flow chart of FIG. 7, it is to be understood that any number of values such of 5 may be used.
The flow chart of FIG. 8 illustrates another arrangement for estimating the prescription product sales volume of a prescribing physician. According to the arrangement of FIG. 8, a group of sampled pharmacies is selected for each unsampled pharmacy and a weighting factor for each sampled pharmacy in the neighborhood of one of the unsampled pharmacies is generated as in Fig. 6. The weighting factors for the sampled pharmacies are combined with the actual sales data for the sampled pharmacies according to
which corresponds to
Vτj = ΣjVSij[l + Σ4Wip]
where Vsij is the prescription product sales volume for physician j at pharmacy i, wφ is the weighting factor for a sampled pharmacy i in the selected neighborhood of unsampled pharmacy p, Σ, is the summation over all sampled pharmacies and Σ2 is the summation over all unsampled pharmacies, Σ3 is the summation of all sampled pharmacies in the neighborhood of unsampled pharmacy p and r4 is the summation of weighting factors associated with sampled pharmacy i. The resulting estimate of sales volume for the prescribing physician is similar to that described with respect to FIG. 6 but the efficiency of the estimate generation is improved.
In the method of FIG. 8, the sampled pharmacies in the neighborhood of each sampled pharmacy are first determined and the weighting factor signals wιp for the neighborhood sampled pharmacies are formed in the processor 215 of FIG. 2 on the basis of the pharmacy location, the characteristics data and the physician identification data in store 205. The sales data from the sampled pharmacies and the weighting factor signals wφ are transferred to the transfer store 228, divided therein into prescribed ranges and each range of data is supplied to an assigned one of the work station processor 230-1 through 230-N. In each work station processor, the estimated sales volume for each sampled pharmacy outlet in range processed by the work station processor is generated from the sales data and the projection factor for the sampled pharmacy outlet. The total estimated sales volume for the sampled pharmacy outlets over range of the work station processor is then generated. When the work station processing is completed the resulting estimated sales volumes from the work station processors are merged and transferred via the transfer store 228 to the processor 215.
Referring to FIG. 8, the locations, type characteristics of pharmacy outlets 01 through ON in FIG. 1 and physician identification data are stored in the characteristics and locations store 205 of FIG. 2 in step 801. Prescription data is received from the pharmacies 01, 02, 03, ON-1 and ON and is stored in data store 210 according to the prescribing physicians (step 803) . The total sales for the prescribing physician j of a particular prescription product is estimated in steps 805 through 870. In step 805, the processor 215 operates to identify the sampled pharmacy outlets Sij and the unsampled pharmacy outlets Up for a particular prescription product according to the validity and volume of the transferred prescription data of the prescribing physician j. The arrangement shown in FIG. 4 may be used. As described with respect to FIGS. 1 and 3, pharmacy outlets 01, 02, 03, ON-l and ON can be determined as sampled outlets Sij, S2j , S3j, S4j and S5j where the sampled data is validated. Outlets 04 and 05 are classified as unsampled outlets Ul and U2.
Step 810 is entered from step 805 and signals representative of the distances dip between each sampled outlet Sij and each unsampled outlet Up are generated in the processor 215. After the distance determination of step 810, the set of nearest sampled pharmacy outlets R j for each unsampled pharmacy outlet Up is selected by processor 215 according to step 815. The selection of sampled pharmacies Rmj may be performed as previously described with respect to FIG. 5.
The unsampled pharmacy index p is reset to one in step 820 and weighting signals wmp are formed for the selected sampled pharmacies Rmi in the loop from step 825 to 834. In step 825, the weighting signal
is formed for each sampled pharmacy m selected as in the neighborhood of sampled pharmacy Up. After the weighting signals for all unsampled pharmacies are formed in the processor 215, the sales data stored in store 210 and the weighting signals are transferred to transfer store 228 wherein the data and weighting signal array is divided into N portions and each portion is transferred to one of work station processors 230-1 through 230-N. The processing of the portion transferred to work station processor 230-1 is shown from step 838-1 through 855-1 in FIG. 8 and the processing the portion transferred to work station processor 230-N is indicated from step 838-1 to 838-N.
With respect to the processing of the range of sampled pharmacies in work station processor 230-1, the sampled pharmacy index i is set to one in step 838-1 and the loop from step 840-1 through 850-1 is iterated to form the estimated prescription volume V'ϋ for the sampled pharmacies in the range from i=l to i=il. In each iteration, an estimated volume signal is generated for sampled pharmacy i in step 840 according to
VTSij = Vsij [l + ∑w^]
where Vsij is the actual prescription product sales volume for physician j at pharmacy i, w^ are the weighting factor for sampled pharmacy Sij and Σ is the summation over all weighting factors associated with the pharmacy Sij. [1 + ∑w,^] is a projection factor for a physician's prescription at the sampled pharmacy VSlj.
After the estimated volume signal is formed for the range i=l to il in the work station processor 230-1, step 855-1 is entered from step 850-1 wherein the total volume for the range from i=l to i=il by summing the. estimated volumes for the pharmacies Sij . The work station processor 230-N operates in similar manner from step 838-N to 855-N to generate a total estimated volume for the range from iN to imax. Work station processors are interconnected by the network 250 such as an ethernet or token ring arrangement so that signals from one work station processor that are required for the formation of the VTSij signal in another work station are transferred. The total volume signals for the ranges are merged in step 860 and the merged signals are transferred to the processor 215 via the transfer store 228. The resulting estimated total volume signal Vτj for the physician j is then formed in the processor 215 (step 865) and a confidence signal for the estimated total volume VTj is generated in step 870 as described with respect to the flow chart of FIG. 7.
The invention has been described with respect to particular illustrative embodiments. It is to be understood that the present invention is not limited to the above described embodiments and that various changes and modifications may be made by those of ordinary skill in the art without departing from the spirit and scope of the invention.

Claims

WHAT IS CLAIMED IS:
1. A method for estimating sales activity of a product at sales outlets including a plurality of sales outlets that transfer sales data to a central station and at least one other sales outlet comprising the steps of: determining a distance between each of the plurality of sales outlets and the at least one other sales outlet; and forming an estimate of the sales activity of the product by correlating the sales data transferred from the plurality of sales outlets to the central station according to the determined distances.
2. A method for estimating sales of a product according to Claim 1 further comprising the step of: forming a signal characterizing each of the plurality of sales outlets and the at least one other sales outlet; and wherein the forming of the estimate of sales of the product includes correlating the transferred sales data according to the determined distances and the characterizing signal for each of the plurality of sales outlets and the at least one other sales outlet.
3. A method for estimating the sales of a product according to Claim 2 , wherein the correlation of the transferred sales is performed according to
vυ = ∑{ { (1/dlu) / (ΣTs/dg }*TU}VS
where Vu is the estimated volume of sales of the product at the at least one other sales outlet, Vs is volume of sales at each of the plurality of sales outlets, Ts is a parameter characterizing each of the plurality of sales outlets, Tυ a parameter characterizing the at least one other'sales outlet, q is greater than zero and the summation Σ is over the plurality of sales outlets.
4. A method for estimating the sales of a product according to Claim 3, wherein the characterizing signal is representative of the size of each of the plurality of sales outlets and the at least one other sales outlet.
5. A method for estimating the sales of a product according to Claim 3, wherein the characterizing parameter signal is representative of the total sales volume of each of the plurality of sales outlets and the at least one other sales outlet.
6. A method for estimating sales of a product at sales outlets including a plurality of sales outlets that transfer sales data to a central station and a first sales outlet comprising the steps of: determining a distance between the first sales outlet and each of the plurality of sales outlets; and forming an estimate of the sales of the product at the first sales outlet by correlating the sales data transferred from the plurality of sales outlets to the central station according to the determined distances.
7. A method for estimating sales of a product according to Claim 6 further comprising the step of: forming a signal characterizing the first sales outlet and each of the plurality of sales outlets; and wherein the forming of the estimate of sales of the product includes correlating the transferred sales data according to the determined distances and the characterizing signal for each of the first and the plurality of sales outlets.
8. A method for estimating the sales of a product according to Claim 7, wherein the correlation of the transferred sales is performed according to
V0 = ∑{{(l/d§u)/(ΣTs/dlu)}*Tu}Vs
where Vυ is the estimated volume of sales of the product at the first sales outlet, Vs is volume of sales at each of the plurality of sales outlets, Ts is a parameter characterizing each of the plurality of sales outlets, Tu a parameter characterizing the first sales outlet, q is greater than zero and the summation Σ is over the plurality of sales outlets.
9. A method for estimating the sales of a product according to Claim 8, wherein the characterizing signal is representative of the size of each of the first and the plurality of sales outlets.
10. A method for estimating the sales of a product according to Claim 8, wherein the characterizing parameter signal is representative of the total sales volume of each of the first and the plurality of sales outlets.
11. A method of estimating sales activity of a controlled product controlled by a controlling authority at sales locations including a plurality of first sales locations at which sales data of controlled products are sampled and at least one second sales location comprising the steps of: generating a signal representing a distance between the at least one second sales location and each of the first sales locations at a central station; transferring controlled product sales data from each first sales location to the central station; selecting a predetermined number of first locations at said central station; and forming a signal representing an estimate of the controlled product sales at the at least one second sales location responsive to the controlled product sales data of the controlling authority at the selected first locations and the distance signals for the selected first sales locations at said central station.
12. The method according to Claim 11 further comprising the step of: forming a signal characterizing the plurality of first sales locations and the at least one second location; and wherein the forming of the estimate of sales of the controlled product includes correlating the transferred sales data according to the determined distances and the characterizing signal for each of the selected first sales locations and the at least one second sales location.
13. The method according to Claim 12, wherein the correlation of the transferred sales of the controlled product is performed according to
vυi = r{{(i/d3u)/(∑τs/_2u)}*τu}vs
where V^ is the estimated volume of sales of the controlled product at the at least one second sales location for the controlling authority j , Vs is volume of sales of the controlled product at each of the selected first sales locations for the controlling authority j , Ts is a parameter characterizing each of the selected first sales locations, Ty a parameter characterizing the at least one second sales location, q is greater than zero and the summation Σ is over the selected first sales locations.
14. The method according to Claim 13, wherein the characterizing signal is representative of the size of each of the selected first sales locations and the at least one second sales location.
15. The method according to Claim 13, wherein the characterizing parameter signal is representative of the total sales volume of each of the selected first sales locations and the at least one second sales location.
16. A method for estimating prescription sales of a prescription product for a prescribing physician at prescription sales locations including a plurality of prescription sales locations which transfer prescription sales information to a central station and a first prescription sales location comprising the steps of: determining a distance between the first prescription sales location and each of the plurality of prescription sales locations transferring sales information to the central station; and correlating the prescription sales information of the prescribing physician from the plurality of prescription sales locations transferring prescription sales information of the prescribing physician to the central station according to the determined distances to form a signal representing an estimate of the prescription sales for the prescribing physician at the first location.
17. The method according to Claim 16 further comprising the step of: forming a signal characterizing the first prescription sales location and each of the plurality of prescription sales locations transferring sales information to the central station; and wherein the forming of the estimate of sales of the prescription product for the physician includes correlating the transferred sales information according to the determined distances and the characterizing signal for each of the plurality of prescription sales locations transferring sales information to the central station and the first prescription sales location.
18. The method according to Claim 17, wherein the correlation of the transferred sales of the prescription product is performed according to
Vυi = ∑{{(l/dgu)/(ΣTs/d§u)}*Tu}Vs
where Vυi is the estimated volume of sales of the prescription product for the prescribing physician j at the first prescription sales location, Vs is volume of sales of the prescription product at each of the plurality of prescription sales locations transferring sales information of the prescribing physician j to the central station, Ts is a parameter characterizing each of the prescription sales locations transferring sales information of the prescribing physician to the central station, Tu a parameter characterizing the first prescription sales location, q is greater than zero and the summation Σ is over the prescription sales locations transferring sales information of the prescribing physician to the central station.
19. The method according to Claim 18, wherein the characterizing signal is representative of the size of each of the prescription sales locations transferring sales information of the prescribing physician to the central station and the first prescription sales location.
20. The method according to Claim 18, wherein the characterizing parameter signal is representative of the total sales volume of each of the prescription sales locations transferring sales information to the central station and the first prescription sales location.
21. the method according to Claim 16 further comprising the step of: selecting prescription sales locations in a predetermined neighborhood of the first prescription sales location responsive to the determined distances; and wherein the correlating of the prescription sales information of the prescribing physician from the plurality of prescription sales locations transferring prescription sales information of the prescribing physician to the central station is performed according to the determined distances for the selected prescription sales locations.
22. A method for estimating prescription sales of a prescription product prescribed by a prescribing physician at prescription sales locations including a plurality of first prescription sales locations which transfer prescription product sales information signals to a central station and a plurality of second prescription sales locations comprising the steps of: determining a distance between the each second prescription sales location and each first prescription sales location; selecting a group of the first prescription sales locations for each second prescription sales location responsive to the determined distances; receiving sales information signals from the first sales locations; correlating the prescription product sales information signals of the prescribing physician from the selected group of first prescription sales locations for each second prescription sales location according to the determined distances between the second prescription sales location and the selected group of the first prescription sales locations for the second prescription sales location to form a signal representing an estimate of the sales of the prescription product for the prescribing physician at the second prescription sales location; forming a signal representing the prescription product sales of the prescribing physician at each of the first prescription sales locations responsive to the prescription sales information signals received by the central station; and combining the signal representative of the prescription product sales of the prescribing physician at the first prescription sales locations with the signal representing the prescription product estimate signal for the prescribing physician at the second prescription sales locations to form a signal representative of an estimate of the prescription product sales for the prescribing physician.
23. The method according to Claim 22 further comprising the step of: forming a signal characterizing each first prescription sales location and each second prescription sales locations; and wherein the forming of the estimate of sales of the prescription product for the physician includes correlating the transferred sales information signals according to the determined distances and the characterizing signal for second prescription sales location and each of the selected group of first prescription sales locations for the second prescription sales location.
24. The method according to Claim 23, wherein the correlation of the transferred sales of the prescription product is performed according to
V0j = ∑ { { ( 1/dSu) / (∑Ts/d§u) } *Tu}VSj
where Vυi is the estimated volume of sales of the prescription product for the prescribing physician j at the first prescription sales location, Vs is volume of sales of the prescription product at each of the plurality of prescription sales locations transferring sales information of the prescribing physician j to the central station, Ts is a parameter characterizing each of the prescription sales locations transferring sales information signals of the prescribing physician to the central station, Tυ a parameter characterizing the first prescription sales location, q is greater than zero and the summation Σ is over the prescription sales locations transferring sales information signals of the prescribing physician to the central station.
25. The method according to Claim 24, wherein the characterizing signal is representative of the size of each first prescription sales locations and each second prescription sales location.
26. The method according to Claim 24, wherein the characterizing parameter signal is representative of the sales volume of all products of each first prescription sales location and each second prescription sales location.
27. A method for estimating prescription sales of a prescription product prescribed by a prescribing physician at prescription sales locations including a plurality of first prescription sales locations which transfer prescription product sales information signals to a central station and a plurality of second prescription sales locations comprising the steps of: determining a distance between the each second prescription sales location and each first prescription sales location; selecting first prescription sales locations for each second prescription sales location responsive to the determined distances; receiving sales information signals from the first prescription sales locations; and correlating the prescription product sales information signals of the prescribing physician from each first prescription sales location according to the determined distances between each first prescription sales location and the second prescription sales locations for which the first prescription sales location is selected to form a signal representing an estimate of the sales of the prescription product for the prescribing physician at the first prescription sales locations.
28. The method according to Claim 27 further comprising the step of: forming a signal characterizing each first prescription sales location and each second prescription sales locations; and wherein the forming of the estimate of sales of the prescription product for the physician includes correlating the transferred sales information signals according to the determined distances and the characterizing signal for each first prescription sales location and each second prescription sales locations for which the first prescription sales location is selected.
29. The method according to Claim 28, wherein the correlation of the received sales of the prescription product is performed according to
VTj = Σ,VSij [l + Σ2w,
where VTj is the total estimated sales for the prescribing physician j , Vsij is the prescription product sales volume for physician j at pharmacy i, wsu is the weighting factor for a sampled pharmacy S in the selected neighborhood of the unsampled pharmacy U, Σ, is the summation over all sampled pharmacies and Σ2 is the summation over the weighting factors associated with sampled pharmacies S and
wsu = { (l/d,2 ) / (Σ3Ts/d,«u ) } *Tυ
where dβu is the distance between a selected first prescription sales outlet and one of the second prescription sales outlets, Ts is a parameter characterizing the first prescription sales outlet, Tu is a parameter characterizing the second prescription sales outlet and q is a number greater than zero and Σ3 is the summation over sampled pharmacies S in a neighborhood of unsampled pharmacy U.
30. The method according to Claim 29 wherein the characterizing signal is representative of the size of each first prescription sales location and each second prescription sales location.
31. The method according to Claim 29, wherein the characterizing parameter signal is representative of the sales volume of all products of each first prescription sales location and each second prescription sales location.
32. In apparatus having a first processing system and a plurality of second processing systems, a method for estimating sales volume of a controlled product authorized by a controlling authority at sales locations including a plurality of first sales locations and a plurality of second sales locations at which sales data of products controlled by the controlling authority are sampled comprising the steps of: generating a set of signals in said first processing system each representing a distance between each first sales location and each second sales location in said first processor; generating a signal for each first sales location and for each second sales location for characterizing the sales location in said first processing system; selecting a group of the second locations for each first location responsive to the distance and characterizing signals in said first processing system; generating a set of signals in said first processing system each representative of the effect of one of the group of second sales locations on the sales of the controlled product at the first sales location for which the second sales location is selected responsive to the distance and characterizing signals; receiving controlled product sales data from the plurality of second locations in the first processing system; transferring the set of effect signals and the transferred controlled product sales data from the first processing system to the plurality of second processing systems; and forming a signal in said plurality of second processing systems representing an estimate of the controlled product sales at the first locations responsive to the effect signals and the transferred controlled product sales data from the first processing system.
33. In apparatus having a first processing system and a plurality of second processing systems, the method according to Claim 32 wherein the system further comprises a store for transferring signals between said first processing system and the plurality of second processing systems and wherein the controlled product sales data and the effect signals are transferred to the transfer store and a portion of the controlled product sales data and the effect signals is transferred from the transfer store to each of the plurality of second processing systems.
34. In apparatus having first and second processing systems, the method according to Claim 32, wherein the sales locations are pharmacy sales outlets and the controlling authority is a prescription prescriber.
35. In apparatus having a first processing system and a plurality of second processing systems, a method for estimating sales volume of a controlled product authorized by a controlling authority at sales locations including a plurality of first sales locations at which sales data of products controlled by the controlling authority are sampled and a plurality of second sales locations comprising the steps of: generating a set of signals in said first processing system each representing a distance between each of the first locations and each of the plurality of second locations; generating a signal for each first sales location and for each second sales location for characterizing the sales location in said first processing system; selecting a group of the first sales locations for each second sales location responsive to the distance signals in said first processing system; receiving controlled product sales data from the plurality of first locations in the first processing system; generating a set of signals in said first processing system representative of the effect of each second sales location on the sales of the controlled product at the first sales location selected for the second sales location responsive to the distance and characterizing signals;
transferring the set of effect signals and the transferred controlled product sales data from the first processing system to the plurality of second processing systems; and forming a signal in said plurality of second processing systems representing an estimate of the controlled product sales at the first locations responsive to the effect signals and the transferred controlled product sales data from the first processing system.
36. In apparatus having a first processing system and a plurality of second processing systems, the method according to Claim 35 wherein the system further comprises a store for transferring signals between said first processing system and the plurality of second processing systems and wherein the controlled product sales data and the effect signals are transferred to the transfer store and a portion of the controlled product sales data and the effect signals is transferred from the transfer store to each of the plurality of second processing systems.
37. The method according to Claim 35 further comprising the step of merging the estimated sales volume signals for the first sales locations in the plurality of second processing systems to produce a signal representative of an estimate of the total sales of the controlling authority.
38. In apparatus having first and second processing systems, the method according to Claim 37, wherein the sales locations are pharmacy sales outlets and the controlling authority is a prescription prescriber.
39. In apparatus having first and second processing systems, the method according to Claim 35, wherein the sales locations are pharmacy sales outlets and the controlling authority is a prescription prescriber.
40. A system for estimating sales activity of a product at sales outlets including a plurality of sales outlets at which sales activity data is sampled and at least one other sales outlet comprising: means for receiving sales data from each of the plurality of sales outlets; first processing means for generating a plurality of signals each representing a distance between one of the plurality of sales outlets and the at least one sales outlet and for generating a plurality of signals each characterizing one of the plurality of sales outlets and the at least one sales outlet; means for dividing the received sales activity data, the distance signals and the characterizing signals into portions each related to a group of the plurality of sales outlets; a plurality of second processing means each for receiving one of the portions for forming an estimate of the sales activity of the product at the group of the plurality of sales outlets by correlating the received sales data according to the distance signals and the characterizing signals.
EP94912949A 1993-04-05 1994-04-05 Method of estimating product distribution Ceased EP0695445A4 (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
EP06023574A EP1777654A3 (en) 1993-04-05 1994-04-05 Method of estimating product distribution

Applications Claiming Priority (3)

Application Number Priority Date Filing Date Title
US08/042,518 US5420786A (en) 1993-04-05 1993-04-05 Method of estimating product distribution
US42518 1993-04-05
PCT/US1994/003681 WO1994023382A1 (en) 1993-04-05 1994-04-05 Method of estimating product distribution

Related Child Applications (1)

Application Number Title Priority Date Filing Date
EP06023574A Division EP1777654A3 (en) 1993-04-05 1994-04-05 Method of estimating product distribution

Publications (2)

Publication Number Publication Date
EP0695445A1 true EP0695445A1 (en) 1996-02-07
EP0695445A4 EP0695445A4 (en) 1997-06-11

Family

ID=21922363

Family Applications (2)

Application Number Title Priority Date Filing Date
EP94912949A Ceased EP0695445A4 (en) 1993-04-05 1994-04-05 Method of estimating product distribution
EP06023574A Withdrawn EP1777654A3 (en) 1993-04-05 1994-04-05 Method of estimating product distribution

Family Applications After (1)

Application Number Title Priority Date Filing Date
EP06023574A Withdrawn EP1777654A3 (en) 1993-04-05 1994-04-05 Method of estimating product distribution

Country Status (6)

Country Link
US (2) US5420786A (en)
EP (2) EP0695445A4 (en)
JP (1) JP3560611B2 (en)
AU (1) AU6530194A (en)
CA (1) CA2159740A1 (en)
WO (1) WO1994023382A1 (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2008141679A1 (en) * 2007-05-21 2008-11-27 Wincor Nixdorf International Gmbh Method and device for determining a contact time for contacting a money transfer system

Families Citing this family (68)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5873069A (en) * 1995-10-13 1999-02-16 American Tv & Appliance Of Madison, Inc. System and method for automatic updating and display of retail prices
JPH1097671A (en) * 1996-09-20 1998-04-14 Media Maaketeingu Network:Kk Commodity sales management method and device for automatic vending machine
US6061657A (en) * 1998-02-18 2000-05-09 Iameter, Incorporated Techniques for estimating charges of delivering healthcare services that take complicating factors into account
US6842736B1 (en) 1998-10-21 2005-01-11 David J. Brzozowski Drug auditing method and system
US7512542B1 (en) * 1999-06-10 2009-03-31 A.C. Nielsen (Us), Inc. Method and system for market research data mining
US6810384B1 (en) * 1999-08-04 2004-10-26 General Mills, Inc. Method and apparatus for sales volume and share decomposition
AUPQ206399A0 (en) * 1999-08-06 1999-08-26 Imr Worldwide Pty Ltd. Network user measurement system and method
US6304854B1 (en) 1999-09-09 2001-10-16 Dunhill Holdings, Corp. System and method for providing a comparable branded product based on a current branded product for non-comparison shopped products
AU7596500A (en) 1999-09-20 2001-04-24 Quintiles Transnational Corporation System and method for analyzing de-identified health care data
US6636837B1 (en) * 2000-01-27 2003-10-21 Eastman Kodak Company Method and apparatus for ordering photofinishing goods and/or services
JP4518357B2 (en) * 2000-11-02 2010-08-04 株式会社イーエムシステムズ Pharmacy support system
US7580851B1 (en) 2000-11-16 2009-08-25 Ims Software Services Ltd. System and method for estimating daily sales volume
US7174304B1 (en) * 2000-12-05 2007-02-06 Ims Health Incorporated System and method for estimating product distribution using a product specific universe
WO2002049257A2 (en) * 2000-12-12 2002-06-20 Nintendo Of America Inc. System for identifying unauthorized sales and providing post-sale information to purchasers using electronic product registration
US7080052B2 (en) 2001-04-19 2006-07-18 International Business Machines Corporation Method and system for sample data selection to test and train predictive algorithms of customer behavior
WO2003001428A2 (en) * 2001-05-04 2003-01-03 Ims Health Incorporated System and methods for estimating product sales in highly fragmented geographical segments of service provider location
FR2831974A1 (en) * 2001-11-02 2003-05-09 Ims Health System for determining the efficiency of commercial actions undertaken by pharmaceutical laboratories in which prescriber data and information is broken down into geographical units that are based on existing geographical areas
EP1464020A4 (en) * 2002-01-11 2005-03-23 Ims Health Inc Method and system for rapidly projecting and forecasting pharmaceutical market information
US8521619B2 (en) * 2002-03-27 2013-08-27 Autotrader.Com, Inc. Computer-based system and method for determining a quantitative scarcity index value based on online computer search activities
US20050198967A1 (en) * 2002-09-23 2005-09-15 Siemens Westinghouse Power Corp. Smart component for use in an operating environment
WO2004034217A2 (en) * 2002-10-08 2004-04-22 Omnicare, Inc. System for assimilating and using pharmacy data
US7069233B2 (en) * 2002-12-31 2006-06-27 Bracken Foster & Associates, Llc System and method for identification and notification of elevated over-the-counter medication sales with response coordination
US7617115B2 (en) * 2003-02-11 2009-11-10 Cerner Innovation, Inc. System and method for risk-adjusting indicators of access and utilization based on metrics of distance and time
WO2005026884A2 (en) * 2003-09-05 2005-03-24 Ims Health Incorporated Techniques for estimating sales of items through a particular channel
IL161263A0 (en) * 2004-04-02 2004-09-27 Crossix Solutions Llc A privacy preserving data-mining protocol
US20050256742A1 (en) * 2004-05-05 2005-11-17 Kohan Mark E Data encryption applications for multi-source longitudinal patient-level data integration
JP2007536833A (en) * 2004-05-05 2007-12-13 アイエムエス ソフトウェア サービシズ リミテッド Multi-source long-term patient-level data encryption
US20050256741A1 (en) * 2004-05-05 2005-11-17 Kohan Mark E Mediated data encryption for longitudinal patient level databases
AU2005241559A1 (en) * 2004-05-05 2005-11-17 Ims Software Services, Ltd. Data record matching algorithms for longitudinal patient level databases
US8742944B2 (en) * 2004-06-21 2014-06-03 Siemens Energy, Inc. Apparatus and method of monitoring operating parameters of a gas turbine
CA2575310C (en) * 2004-07-28 2014-11-04 Ims Health Incorporated A method for linking de-identified patients using encrypted and unencrypted demographic and healthcare information from multiple data sources
US20060190288A1 (en) * 2005-01-22 2006-08-24 Ims Software Services Ltd. System and method for allocating prescriptions to non-reporting outlets
US8078488B2 (en) * 2005-01-25 2011-12-13 Ims Software Services Ltd. System and method for determining trailing data adjustment factors
US8744897B2 (en) * 2005-01-22 2014-06-03 Ims Software Services Ltd. Sample store forecasting process and system
US7813953B2 (en) * 2005-01-22 2010-10-12 Bank Of America, N.A. System and method for product level projections of pharmacy prescriptions within product therapy classes
US7921029B2 (en) * 2005-01-22 2011-04-05 Ims Software Services Ltd. Projection factors for forecasting product demand
US8498891B2 (en) * 2005-01-22 2013-07-30 Ims Software Services Ltd. System and method for product level projections of pharmacy prescriptions within product therapy classes
US7314170B2 (en) * 2005-01-25 2008-01-01 Ims Software Services Ltd. System and method for product imputation relating to sample stores
WO2006080991A2 (en) * 2005-01-25 2006-08-03 Ims Software Services, Inc. System and method for allocating prescriptions to non-reporting outlets
CA2595352A1 (en) * 2005-01-25 2006-08-03 Ims Software Services Ltd. Sample store forecasting process and system
WO2007102794A2 (en) * 2005-01-25 2007-09-13 Ims Software Services Projection factors for forecasting product demand
US8103539B2 (en) * 2005-01-25 2012-01-24 Ims Software Services Ltd. Sample store forecasting process and system
EP1844414A4 (en) * 2005-01-25 2010-02-10 Ims Software Services Inc System and method for product level projections of pharmacy prescriptions within product therapy classes
US8577933B2 (en) * 2006-08-02 2013-11-05 Crossix Solutions Inc. Double blinded privacy-safe distributed data mining protocol
US9355273B2 (en) 2006-12-18 2016-05-31 Bank Of America, N.A., As Collateral Agent System and method for the protection and de-identification of health care data
US20080262900A1 (en) * 2007-04-18 2008-10-23 Michael Day Duffy Methods and apparatus to facilitate sales estimates
WO2008157287A2 (en) * 2007-06-14 2008-12-24 The Nielsen Company (U.S.), Inc. Methods and apparatus to weight incomplete respondent data
CA2691548A1 (en) * 2007-06-29 2009-01-08 Ims Software Services, Ltd. Systems and methods for projecting sample store activities that are restricted in non-sample stores
US20090319331A1 (en) * 2008-06-19 2009-12-24 Michael Day Duffy Methods and apparatus to respond to recalls
US20100131284A1 (en) * 2008-11-26 2010-05-27 Michael Day Duffy Methods and apparatus for analysis of healthcare markets
US20100205034A1 (en) * 2009-02-09 2010-08-12 William Kelly Zimmerman Methods and apparatus to model consumer awareness for changing products in a consumer purchase model
US20100306028A1 (en) * 2009-06-02 2010-12-02 Wagner John G Methods and apparatus to model with ghost groups
WO2011035298A2 (en) * 2009-09-21 2011-03-24 The Nielsen Company (Us) Llc Methods and apparatus to perform choice modeling with substitutability data
US20110161108A1 (en) * 2009-12-30 2011-06-30 Mckesson Automation Inc. Systems and methods for detecting diversion in drug dispensing
US20120259676A1 (en) 2011-04-07 2012-10-11 Wagner John G Methods and apparatus to model consumer choice sourcing
US9311383B1 (en) 2012-01-13 2016-04-12 The Nielsen Company (Us), Llc Optimal solution identification system and method
WO2014152010A1 (en) 2013-03-15 2014-09-25 Affinnova, Inc. Method and apparatus for interactive evolutionary algorithms with respondent directed breeding
WO2014143729A1 (en) 2013-03-15 2014-09-18 Affinnova, Inc. Method and apparatus for interactive evolutionary optimization of concepts
US9219928B2 (en) 2013-06-25 2015-12-22 The Nielsen Company (Us), Llc Methods and apparatus to characterize households with media meter data
US9277265B2 (en) 2014-02-11 2016-03-01 The Nielsen Company (Us), Llc Methods and apparatus to calculate video-on-demand and dynamically inserted advertisement viewing probability
JP5724020B1 (en) * 2014-06-18 2015-05-27 アイエムエス ソフトウェア サービシズ リミテッド Information processing apparatus and program
US20160034931A1 (en) * 2014-07-31 2016-02-04 Applied Predictive Technologies, Inc. Systems and methods for generating a location specific index of economic activity
US20160086115A1 (en) * 2014-09-18 2016-03-24 Ims Health Incorporated Performance Management by Indication
US10219039B2 (en) 2015-03-09 2019-02-26 The Nielsen Company (Us), Llc Methods and apparatus to assign viewers to media meter data
US10147108B2 (en) 2015-04-02 2018-12-04 The Nielsen Company (Us), Llc Methods and apparatus to identify affinity between segment attributes and product characteristics
US10791355B2 (en) 2016-12-20 2020-09-29 The Nielsen Company (Us), Llc Methods and apparatus to determine probabilistic media viewing metrics
CA3066633A1 (en) * 2017-06-30 2019-01-03 Christian Lafrance System and method for exposing and integrating multiple supply chain and delivery networks to optimize capacity utilizations
US11074548B2 (en) * 2019-12-05 2021-07-27 Coupang Corp. Computer implemented systems and methods for optimization of a product inventory by intelligent distribution of inbound products

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US4926394A (en) * 1989-10-27 1990-05-15 Western Atlas International, Inc. Monte Carlo method for estimating lithology from seismic data
JPH03224054A (en) * 1990-01-30 1991-10-03 Toshiba Corp Sales amount estimating device
GB2250362A (en) * 1990-05-07 1992-06-03 Mitsubishi Electric Corp Parallel data processing system
US5377095A (en) * 1991-07-12 1994-12-27 Hitachi, Ltd. Merchandise analysis system with sales data table and various functions for predicting the sale by item

Family Cites Families (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US4972504A (en) * 1988-02-11 1990-11-20 A. C. Nielsen Company Marketing research system and method for obtaining retail data on a real time basis
US4975843A (en) * 1988-11-25 1990-12-04 Picker International, Inc. Parallel array processor with interconnected functions for image processing
US5063519A (en) * 1989-09-18 1991-11-05 Pacific Energy Landfill gas production testing and extraction method
US4995011A (en) * 1989-09-20 1991-02-19 Woods Hole Oceanographic Institute Acoustic mapping system using tomographic reconstruction
US5198979A (en) * 1989-09-26 1993-03-30 Shell Oil Company Seismic migration of multiprocessor computer

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US4926394A (en) * 1989-10-27 1990-05-15 Western Atlas International, Inc. Monte Carlo method for estimating lithology from seismic data
JPH03224054A (en) * 1990-01-30 1991-10-03 Toshiba Corp Sales amount estimating device
GB2250362A (en) * 1990-05-07 1992-06-03 Mitsubishi Electric Corp Parallel data processing system
US5377095A (en) * 1991-07-12 1994-12-27 Hitachi, Ltd. Merchandise analysis system with sales data table and various functions for predicting the sale by item

Non-Patent Citations (5)

* Cited by examiner, † Cited by third party
Title
CHONG D.: 'BEYOND A DEMOGRAPHIC INFORMATION SYSTEM: RETAIL AND SERVICE INDUSTRY APPLICATIONS OF GEOGRAPHIC INFORMATION SYSTEMS' GIS/LIS PROCEEDINGS 10 November 1992 - 12 November 1992, SAN JOSE, USA, pages 121 - 128 *
CHOU YUE-HONG: 'analyzing the spatial autocorrelation of polygonal data in GIS' URISA PROC. 06 August 1989 - 10 August 1989, BOSTON, MA, USA, pages 138 - 148 *
ISAKI CARY T.: 'JOURNAL OF BUSINESS & ECONOMIC STATISTICS' vol. D, no. 4, October 1990, pages 435 - 441 *
PATENT ABSTRACTS OF JAPAN vol. 015, no. 515 (P-1293), 27 December 1991 & JP 03 224054 A (TOSHIBA CORP), 3 October 1991, *
See also references of WO9423382A1 *

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2008141679A1 (en) * 2007-05-21 2008-11-27 Wincor Nixdorf International Gmbh Method and device for determining a contact time for contacting a money transfer system

Also Published As

Publication number Publication date
WO1994023382A1 (en) 1994-10-13
US5781893A (en) 1998-07-14
EP1777654A2 (en) 2007-04-25
JPH08511885A (en) 1996-12-10
AU6530194A (en) 1994-10-24
EP0695445A4 (en) 1997-06-11
EP1777654A3 (en) 2008-07-16
CA2159740A1 (en) 1994-10-13
US5420786A (en) 1995-05-30
JP3560611B2 (en) 2004-09-02

Similar Documents

Publication Publication Date Title
US5781893A (en) System for estimating product distribution
Johnston et al. Multiscale spatial pattern in nonuse willingness to pay: applications to threatened and endangered marine species
Garber et al. From density to destiny: Using spatial dimension of sales data for early prediction of new product success
Fridgen et al. Management zone analyst (MZA) software for subfield management zone delineation
US8583562B1 (en) Predicting real estate and other transactions
US8265984B2 (en) Method and apparatus for unified optimization model for retail network configuration
Le Ray et al. Online adaptive clustering algorithm for load profiling
US20090187464A1 (en) Method and apparatus for end-to-end retail store site optimization
Alaeddini et al. A hybrid prediction model for no-shows and cancellations of outpatient appointments
Bivand A review of spatial statistical techniques for location studies
US7174304B1 (en) System and method for estimating product distribution using a product specific universe
US8041590B2 (en) In-store media rating system and method
Cromley et al. Evaluating representation and scale error in the maximal covering location problem using GIS and intelligent areal interpolation
US20020178167A1 (en) System and method for allocating data in a hierarchical organization of data
KR20140142691A (en) Market survey and analysis system
Rojas et al. Modeling lot-size with time-dependent demand based on stochastic programming and case study of drug supply in Chile
Frazier et al. Spatial econometric models for panel data: incorporating spatial and temporal data
Monogan et al. Measuring state and district ideology with spatial realignment
Nicoletta et al. Bayesian spatio-temporal modelling and prediction of areal demands for ambulance services
Baker et al. Predicting treatment effect from surrogate endpoints and historical trials: an extrapolation involving probabilities of a binary outcome or survival to a specific time
Anggrainingsih et al. Forecasting The Number of Patients at RSUD Sukoharjo Using Double Exponential Smoothing Holt
Hornberger et al. Optimal heterogeneous search and rescue asset location modeling for expected spatiotemporal demands using historic event data
Jackson et al. Bayesian multistate modelling of incomplete chronic disease burden data
Zhang et al. Dynamic time warp-based clustering: Application of machine learning algorithms to simulation input modelling
Kettani et al. PariTOP: A goal programming-based software for real estate assessment

Legal Events

Date Code Title Description
PUAI Public reference made under article 153(3) epc to a published international application that has entered the european phase

Free format text: ORIGINAL CODE: 0009012

17P Request for examination filed

Effective date: 19951103

AK Designated contracting states

Kind code of ref document: A1

Designated state(s): AT BE CH DE DK ES FR GB GR IT LI NL

A4 Supplementary search report drawn up and despatched
AK Designated contracting states

Kind code of ref document: A4

Designated state(s): AT BE CH DE DK ES FR GB GR IT LI NL

17Q First examination report despatched

Effective date: 19981209

APBN Date of receipt of notice of appeal recorded

Free format text: ORIGINAL CODE: EPIDOSNNOA2E

APBR Date of receipt of statement of grounds of appeal recorded

Free format text: ORIGINAL CODE: EPIDOSNNOA3E

APAA Appeal reference recorded

Free format text: ORIGINAL CODE: EPIDOS REFN

APAF Appeal reference modified

Free format text: ORIGINAL CODE: EPIDOSCREFNE

APBT Appeal procedure closed

Free format text: ORIGINAL CODE: EPIDOSNNOA9E

STAA Information on the status of an ep patent application or granted ep patent

Free format text: STATUS: THE APPLICATION HAS BEEN REFUSED

18R Application refused

Effective date: 20061115